The objective of the project was to develop an approach to use Raman spectroscopy to predict sensory quality of fresh pork loins. Pork loins (n=200) were selected during fabrication on seven different production days and six different production plants. Loins were selected to represent the range of quality features. Scans of the ventral portion of the intact loin were collected with a portable Raman spectrometer. Scan time was 6 seconds. Loins were then packaged and aged 14 days. Pork chops from each loin were cut and packaged 14 days postmortem. Scans of the cross section of central loin chops were collected on all 200 loins. Each chop was scanned in three locations of the cross section with a scan time of 6 seconds. Pork sensory quality was evaluated on a select group of loins (n=75 per group) to determine sensory tenderness, chewiness, juiciness, flavor, and off-flavor. Slice shear force was measured on all loins (n=200 per group).
Raman peaks are represented by their wavenumber (Raman shift) and intensity. The peak intensities are dependent on many factors that may vary from sample to sample (i.e., sample size, exposure time, etc.), but their Raman shifts (i.e., the peak wavelengths) remain identical as long as the molecular makeup is the same. These data are summarized using a barcode approach to highlight unique Raman shift “fingerprints” of each sample. Binary barcodes are generated for each sample based on the second derivative spectra collected. Each sample has a barcode that is specific to its Raman spectral properties. The barcode was used to determine the association with sensory traits and pH traits.
Raman spectral properties of pork loin at day 1 postmortem were approximately 60-70 % accurate in predicting aged pork loin tenderness traits. Raman properties of each pork loin after aging were generally over 90 % accurate. Prediction of cooked pork loin juiciness was better than that for tenderness traits.
Aged pork pH was accurately predicted with a regression curve of 11 Raman peaks (within 0.17 pH units). Inclusion of all the spectral data resulted in an improvement in the accuracy of pH prediction to within 0.05 units. This finding is significant and shows promise of developing a non-invasive, non-destructive method to measure an important fresh pork quality trait- pH.
Raman spectral properties show promise in prediction of fresh pork quality. Adoption of the procedure will require engineering to allow measurement at line speeds.